Octopamine signaling pathway (i.e., synthesis, activation, inhibition) and experimental design

(A) Octopamine (OA) is synthesized by conversion of tyrosine to tyramine by tyrosine decarboxylase (TDC) and tyramine conversion to OA by tyramine β-hydroxylase (TβH). (B) OA β-2 receptor (Oβ-2R) is one of the most abundant and specific OA-detecting receptors. Upon activation of Oβ-2R, a G-coupled protein receptor, adenylyl cyclase (AC) is activated and increases intracellular cyclic adenylyl monophosphate (cAMP) levels. High cAMP levels induce protein kinase A activity which phosphorylates metabolic enzymes and transcription factors that enhance metabolic activity. (C) An OA receptor antagonist, epinastine (EP), inhibits OA binding and prevents activation of the Oβ-2R pathway. (D) To evaluate relationships between virus infection and the OA response, honey bees were inoculated with either deformed wing virus (DWV) or sacbrood virus (SBV) via injection or mock infected. Bees were exposed to either OA, EP, or both OA and EP via co-injection or feeding, all bees were fed sucrose. (E) Impacts of experimental conditions were quantified in individual bees 72 hours post infection. Panels B-D were produced using BioRender.com.

The flight performance of honey bees infected with DWV and/or SBV is differentially impacted by OA and EP treatment

To evaluate the involvement of the OA pathway on virus infection and/or flight performance, virus- or mock- infected honey bees were exposed to OA, EP, or both OA and EP via co-injection (blue) or feeding (green). Honey bees were experimentally infected with or had preexisting infections of DWV (A-C) or SBV (D-F) in three experiments. Flight distance (log10 m) was evaluated using linear mixed effect models to quantify impacts of DWV and/or SBV infection in the context of OA and/or EP treatment. Each data point represents an individual bee (n = 336); colors indicate treatment (i.e., bees fed only sucrose in pink, bees fed OA and/or EP in green, and bees injected with OA and/or EP in blue) and gray areas represent 95% confidence intervals. Linear mixed models with estimates in Data S4. (A) In bees fed only sucrose, DWV infections negatively impacted flight distance (p = 0.005), whereas DWV infected bees treated with OA flew further distances (i.e., average flight distance was similar to mock infected bees fed sucrose, p = 0.61 and 0.08, respectively). (B) DWV infected bees injected with EP flew shorter distances than DWV infected bees fed only sucrose (p = 0.002). (C) Co-injection of OA and EP negatively impacted the distance DWV infected honey bees flew (p = 0.020) (D) OA-treatment did not enhance flight performance of SBV infected bees (p = 0.69 and 0.60) (E) Honey bees with SBV infections fed or injected with EP flew shorter distances than SBV infected bees fed only sucrose (p = 0.022 and 0.004, respectively). (F) Bees with SBV infections fed or injected with OA and EP resulted in similar flight distances regardless of SBV infection level.

Expression of tyrosine decarboxylase (tdc) and tyramine β-hydroxylase (tβh) across treatments and SBV infection

The relationships between honey bee tyrosine decarboxylase (tdc) and tyramine β-hydroxylase (tβh) expression in the context of virus infection and treatments were evaluated using a linear mixed effect model that included SBV abundance and OA/EP treatments as fixed effects with an interaction. Individual honey bee data plotted by treatment group with either (A) tdc expression or (B) tβh expression. Gene expression was calculated as log2 fold change using the ΔΔCt method relative to housekeeping rpl8 and compared to mock infected bees. SBV abundance was calculated as log10 copies / 2 µg RNA. The background violin plots represent the 95% confidence interval, and the horizontal line in each violin represents the median. SBV abundance is represented by a green color scale and SBV-negative samples with an “X”. When SBV abundance did not impact tdc or tβh expression the dark green points are evenly distributed above and below the control ‘0’ fold change line, whereas unequal distribution indicates an SBV-specific effect; * p<0.05, **p<0.01, and ***p<0.001. (A) Positive association between SBV abundance and tdc expression (p = 0.021). SBV infected bees fed OA or both OA and EP had greater tdc expression levels than uninfected bees fed only sucrose (p = 0.03 and 2×10−10). (B) Expression of tβh was greater in SBV infected bees than uninfected bees (p =0.022). SBV infected bees fed OA had greater tβh expression levels than uninfected bees fed sucrose (p = 2×10−5). SBV infected bees injected with EP or co-injected with OA and EP had lower tβh expression (p = 2×10−5 and 4×10−9).

Relative tdc and tβh expression in honey bees with high DWV or SBV levels

(A) Tyrosine is converted via tyrosine decarboxylase (TDC) to tyramine. Tyramine is a neurohormone that acts as a behavioral antagonist to octopamine (OA), resulting in reduced movement in invertebrates. Tyramine may act as a signaling neurohormone, or tyramine β-hydroxylase (TβH) may convert it to OA. (B) A linear mixed model was used to assess the relative expression of tdc and tβh in honey bees harboring high levels of either deformed wing virus (DWV) or sacbrood virus (SBV) (i.e., 108 virus copies / 2µg RNA) in the context of OA and/or epinastine (EP) injection. Estimates of relative tdc and tβh expression were generated based on data from 336 honey bees. The single bold dash indicates tdc or tβh expression was similar in virus infected bees and mock infected bees and arrows indicate estimates that were higher or lower than controls where each arrow represents fold change of 1.

Transcriptome level comparison of honey bees that were SBV infected, DWV infected, or DWV infected and OA-treated

Gene expression was evaluated in honey bees that were mock infected, SBV infected, DWV infected, or DWV infected bees that were co-injected with OA (DWV+OA) after flight. Differentially expressed genes (DEGs) for each treatment group were identified relative to mock infected bees and compared using a Venn diagram. Full list of DEGs, fold change, and Benjamini-Hochberg corrected significance values are included in SI Table S6.

Pathogen testing of virus inoculum and honey bee samples

(A) Filtered virus inoculums DWV, SBV, and DWV+SBV (labeled A, B, or C, respectively) were tested for other common viruses including acute bee paralysis virus (ABPV), Apis mellifera filamentous virus (AmFV), Andrena bee-associated virus-1 (AnBV), black queen cell virus (BQCV), chronic bee paralysis virus (CBPV), deformed wing virus (DWV), Israeli acute paralysis virus (IAPV), Kashmir bee virus (KBV), Lake Sinai viruses 1-4 (LSV), and sacbrood virus (SBV) using virus-specific polymerase chain reaction (PCR) and analyzed by gel electrophoresis; positive (+) and negative (-) controls. (B) Pathogen diagnostic PCR was performed using pooled mock-infected honey bee cDNA from experiments 1-3 (labeled 1-3) as template determined that bees had preexisting DWV (experiments 1-3) and SBV (experiments 1-3). All samples were tested for DWV and SBV via qPCR (Table S1). All experiments were negative for all non-viral pathogens including Ascosphaera apis (Aa.), Lotmaria passim (Lp.), Melisococcus plutonius (Mp.), Nosema ceranae (Nc.), and Paenibacillus larvae (Pl.); cDNA quality was assessed via amplification of the honey bee housekeeping gene, rpl8. Although no positive control was available for A. apis or P. larvae, the primers were utilized successfully in previous studies. To illustrate this result in a single figure, cDNA from mock-infected individual bee samples from all experiments were pooled by experiment (i.e. n=12 from experiment 1, n=12 from experiment 2, n=12 from experiment 3) and PCR was repeated. The products of pathogen-specific PCRs using pooled cDNA (S), positive (+), and negative control (-) templates were analyzed by agarose gel electrophoresis

Estimated flight duration and speed by virus and OA and/or EP treatment

To visualize the effects of virus infection and octopamine (OA) and/or epinastine (EP) treatment on honey bee (A-D) flight duration (minutes) and (E-H) peak flight speed (km/h), we compared estimated means using experimental data analyzed with linear mixed models (S1 Table S4). Each individual point represents the estimated marginalized means with bars indicating one standard deviation of the mean. Green points represent predictions when bees harbored high SBV levels (i.e., 108 SBV RNA copies / 2 µg RNA) and orange points represent predictions when bees harbored high DWV levels (i.e., 108 DWV RNA copies / 2 µg RNA). Blue points represent estimates for virus free bees. (A) SBV infected bees flew similar durations to uninfected bees, and all treatments resulted in similar or shorter flight durations than those fed sucrose only except OA fed bees, which flew for greater durations. (B) DWV infected bees flew for shorter durations than uninfected bees, but EP and OA treatment resulted in greater flight durations. (C-D) Observed flight durations where each point represents data collected from an individual bee (total n = 336) and the color scale represents virus levels. (E) SBV infected bees flew similar peak speeds than uninfected bees, but any OA or EP treatment resulted in lower peak speeds except for OA+EP injected and OA injected bees, which flew similar distances to those fed only sucrose. (F) DWV infected bees flew at slower speeds than uninfected bees. When DWV infected bees were fed or injected with OA, peak speeds were slightly higher. (G-H) Observed peak flight speeds where each point represents data collected from an individual bee with (G) SBV infections or (H) DWV infections.

DWV infected bees had no relationship with tdc and treatment-specific relationships with tβh

To identify relationships between tyrosine decarboxylase (tdc) and tyramine β-hydroxylase (tβh) expression across virus infection and treatments, we evaluated relationships using a linear mixed effect model (SI Table S4). To visualize the data, we plotted all data points against treatment groups with either (A) tdc expression or (B) tβh expression. The background violin plots represent the 95% confidence interval, the individual data points represent individual bee data, the horizontal line in each violin represents the median for the treatment group. Ranges of DWV infection were included as a colorscale (i.e., darker orange representing higher DWV infection levels). If there is no effect of DWV infection, the darkest points would be evenly distributed above and below the control ‘0’ fold change line. SBV levels were more strongly associated with tdc expression (Fig. 3). (A) There was no difference in tdc expression by DWV abundance (p = 0.83). (B) The expression of tβh was greater in DWV infected bees relative to uninfected bees (p<0.0001, SI Table S4). In addition, DWV infected bees fed OA, injected with EP, or injected with OA and EP exhibited reduced tβh expression (p-values <0.001, SI Table S4).

Expression of tdc, tβh, and oβ-2R are positively correlated

The relationships between the expression of tdc, tβh, and oβ-2R in honey bees were evaluated using a linear mixed effect model that included tdc, tβh, and oβ-2R expression, and SBV abundance as fixed effects and experiment, injection, and treatment (i.e., OA and/or EP) as random effects. This model, which explained 95% of the data, determined that tdc, tβh, and SBV abundance were associated with increased oβ-2R expression (p-values <0.05, SI Table S4). Each data point represents data from an individual honey bee across three experiments (total n = 336; SI Tables S1 and S9).

Comparison of honey bee OA-associated gene expression

To identify potential differences in expression between different treatments and octopamine (OA)-associated genes, expression (as variance stabilized transformed values) between different treatments were evaluated. Specifically, the expression of creb subunits (A and B), protein Kinase A (PKA) subunits, OA receptors (oβ-2R, oβ-3R, oβ-1R, and oα-1R), adenylyl cyclase (ac) and the two precursor enzymes that convert tyrosine to tyramine and tyramine to OA (i.e., tdc and tβh, respectively). Each point represents data from an individual bee. Blue boxplots were mock infected with buffer injections, green boxplots were experimentally-infected with sacbrood virus (SBV), orange boxplots were experimentally infected with deformed wing virus (DWV), and pink boxplots were coinjected with DWV and OA (DWV+OA). Lighter colors indicate treatments that did not fly and darker color boxplots indicate treatments that flew (see Fig. 5). We compared the average expression of OA associated genes across treatment groups and determined that SBV infected bees that did not fly had the greatest differences in OA-associated gene expression relative to all other treatment groups (Complete data available in SI Table S9).